How can we help?

Automations and autonomous workflows can perform business operations without continuous human interaction. Unexpected activity may result from configuration errors, incorrect triggers, integration problems, unsuitable instructions, or authorization issues. Potentially unsafe behavior should be investigated promptly. What are examples of unexpected activity? 1. A workflow runs at an unexpected time. 2. An automation processes the wrong business records. 3. An AI Employee initiates an unintended operation. 4. A workflow uses an unexpected integration. 5. An operation repeats more frequently than intended. 6. A process modifies records outside its expected scope. 7. An automation produces unusual execution results. What should I do first? Step 1: Identify the affected automation, workflow, or business process. Step 2: Determine whether the operation is still running or scheduled to run again. Step 3: If the operation may cause harm, use the available controls to stop, pause, or disable it where supported. Step 4: Review the configured trigger or schedule. Step 5: Review the workflow's instructions and dependencies. Step 6: Check the relevant integration permissions. Step 7: Inspect available execution history and error details. Step 8: Notify an authorized administrator if business information or external systems may be affected. Step 9: Contact Polyosync Support if the activity suggests a security problem. What if the workflow changed external records? Review the external system's available history and recovery options. Do not assume that Polyosync can automatically reverse changes performed through an external integration. What if the workflow was triggered by a scheduled task? Review the scheduling configuration and relevant execution context. Scheduling should determine when an operation runs, not grant additional access. What if the activity involved unauthorized data? Stop the affected operation and report the potential exposure. Do not continue processing unrelated information. How can I prevent recurrence? 1. Review the workflow's purpose. 2. Narrow its operational scope. 3. Verify its triggers. 4. Review integration permissions. 5. Test with representative data. 6. Add appropriate human oversight. 7. Review execution results after changes. Recommended practice Treat unexpected automated actions as operational incidents that require investigation, especially when sensitive data or external systems are involved.